Hello!
I get the situation, what called 'Aggressive Data Scanning' in 'SQL Server 2008 White Paper: Analysis Services Performance Guide'
Do anyone know how to resolve this problem?
-------------------------
I have SQL Server 2008 r2 sp1 x64.
In cube definition i have some Dimensions: products_1, time, store, product_2.
In dimension 'products_1' I have defined 3-level user hierarhy.
In dimension 'products_2' I have defined 5-level usrt hierarhy.
Then I have the MDX-query like this:
SELECT { Distinct(Hierarchize( {[Time].[YearMothDay].[All].&[2013].&[3].&[2013-03-24T00:00:00], [Time].[YearMothDay].[All].&[2013].&[3].&[2013-03-25T00:00:00], ..... [Time].[YearMothDay].[All].&[2013].&[4].&[2013-04-20T00:00:00]} )) * {[Measures].[sales], [Measures].[remain]} } DIMENSION PROPERTIES PARENT_UNIQUE_NAME, CHILDREN_CARDINALITY ON COLUMNS, { Distinct(Hierarchize( {[products].[GroupSubgroupProduct].[All].&[Groupname1].&[Subgroupname1].&[44797], --and 49 products more } )) * {[Stores].[DepartmentStore].[Store].AllMembers} } DIMENSION PROPERTIES PARENT_UNIQUE_NAME, CHILDREN_CARDINALITY ON ROWS FROM ( SELECT ( {[Stores].[Status].[All].&[Opened]} ) ON COLUMNS FROM [SSDW] )
if I use products_1 dimension
and same query, if I use product_2 dimension:
SELECT { Distinct(Hierarchize( {[Time].[YearMothDay].[All].&[2013].&[3].&[2013-03-24T00:00:00], [Time].[YearMothDay].[All].&[2013].&[3].&[2013-03-25T00:00:00], ..... [Time].[YearMothDay].[All].&[2013].&[4].&[2013-04-20T00:00:00]} )) * {[Measures].[sales], [Measures].[remain]} } DIMENSION PROPERTIES PARENT_UNIQUE_NAME, CHILDREN_CARDINALITY ON COLUMNS, { Distinct(Hierarchize( {[products].[Hierarhy1].[All].&[level1].&[level2].&[level3].&[level4].&[level5].&[44797], --and 49 products more } )) * {[Stores].[DepartmentStore].[Store].AllMembers} } DIMENSION PROPERTIES PARENT_UNIQUE_NAME, CHILDREN_CARDINALITY ON ROWS FROM ( SELECT ( {[Stores].[Status].[All].&[Opened]} ) ON COLUMNS FROM [SSDW] )
In profiler I sow (EventClass=QuerySubcuveVerbose; EventSubclass=22-Non-cache data):
Dimension 0 [Time] (* * * 0 0 0) [Dimtime]:* [Monthnumber]:* [Yearnumber]:* [Weeknumber]:0 [Dayofweeknumber]:0 [Iso Year]:0 Dimension 1 [Product_1] (+ 0 0 0 0 0 0 0 0 * * 0 0 0 0 0 0 0) [product]:+ [priseSegment]:0 [Brend]:0 [ТМ]:0 [Manager]:0 [Status]:0 [Prodiser]:0 [Weigted]:0 [someattr]:0 [Group]:* [Subgroup]:* [Discount]:0 [someattr2]:0 [someattr3]:0 [someattr4]:0 [someattr5]:0 [someattr6]:0 [someattr7]:0 Dimension 2 [Stores] (* * 0 - 0) [Store]:* [Department]:* [Model]:0 [Status]:- [ObjectType]:0
Ffor first query
and
Dimension 0 [time] (* * 0 * 0 0) [Dimtime]:* [Yearnumber]:* [Iso Year]:0 [Monthid]:* [Weeknomberid]:0 [Dayofwee Id]:0 Dimension 1 [Product_2] (* 0 0 0 0 0 * * * * *) [product]:* [weigted]:0 [Maneger]:0 [Producer]:0 [TM]:0 [PriceSegment]:0 [Level1]:* [level2]:* [level3]:* [Level4]:* [Level5]:* Dimension 2 [Stores] (* * 0 3 0) [Store]:* [Department]:* [Model]:0 [Status]:[Opened] [ObjectType]:0
The fact portitions are big, and in the second caseI get message 'Have`t memory'. (SSAS was get more then 22 GB).
What I must to do to get only needed data to the cashe (like in first case)?